Welcome to my profile! I am a student developer interested in robotics, machine learning, and software/engineering problems that can solve issues in the real-world.
- 🎓 Student at the North Carolina School of Science and Mathematics.
- 🤖 Member of FRC Team 900 Zebracorns, former Captain of Team 2059 Hitchhikers
- 🧠 Focused on machine learning, backend development, and robotics control systems
- 🏡 Outside of engineering, I enjoy rock climbing and playing sports like pickleball, soccer, ultimate frisbee, and football.
- 📫 Email: aarushjain189@gmail.com
Geospatial optimization engine that ranks where to place river trash interception nets, scoring candidate sites on 28 hydrology, feasibility, and environmental justice parameters across 108,772 cities in 240 countries.
- DEM hydrology pipeline in Python with pysheds: depression filling, D8 flow direction, flow accumulation, stream extraction, Manning's equation velocity estimates
- Dirichlet-perturbed Monte Carlo sensitivity analysis to test ranking stability
- FastAPI backend and Mapbox GL dashboard pulling live waterway geometry from OpenStreetMap, scoring any city in under five seconds
- Validated against Ellerbe Creek in Durham, NC
- 1st place at the 2026 U.S. Stockholm Junior Water Prize, advancing to the international finals in Stockholm
- Site
Offline drone ground station that turns a live FPV video feed into mapped survivor coordinates with a stated uncertainty radius, running fully offline on a fanless MacBook Air M4.
- YOLO11s fine-tuned on the HERIDAL and SARD aerial datasets, exported to Core ML FP16 on the Apple Neural Engine at 63 fps
- Custom NumPy ByteTrack tracker with a Kalman filter to bridge 10 fps detection to 30 fps video
- Monocular ray-to-ground projection in a local ENU frame with a 1,000-sample Monte Carlo that reports a 95th-percentile error radius instead of a false-precision pin, reaching 1.2 m median error near nadir
- Electron, React, and TypeScript coordinator interface with an offline MapLibre map and MGRS/WGS84 readouts
- Designed and 3D printed the airframe carrying the DJI O4, ELRS, and GPS stack
- 1 of 5 projects from 250 selected to present at the MIT Media Lab Day of AI, winning the Most Responsible Design Award; Top 100 at Samsung Solve for Tomorrow 2025; $1,000 Bowman Brockman Fund grant
- Site
Mobile app that uses a custom TensorFlow CNN on Mel spectrogram audio to flag possible signs of sleep apnea from phone recordings.
- Built with Flutter/Dart and TensorFlow
- Trained on university audio datasets with spectrogram preprocessing
- Finalist (Top 9) and 3rd place at the 2025 Ready Set App Challenge
- Devpost
- Finals pitch
Official mobile app for the NC TSA state, serving live schedules, real-time updates, and push notifications to 2,000+ attendees.
- React Native (Expo) and Next.js frontend on a Go, PostgreSQL, and Redis backend
- Migrated from a self-hosted home server to Docker and Kubernetes, sustaining 1.8M requests across multi-day competition weekends
- Automations and an AI agent for conference operations
- Built with @NC-TSA
Work across control, vision, and autonomous systems for competition robots.
- Currently on the programming team for @FRC900, the Zebracorns
- Former Captain and programming member of FRC Team 2059 Hitchhikers
- Experience with swerve and tank drivetrains, PhotonVision and custom vision pipelines, PID/feedforward control, motion profiling, CAN bus wiring, electronics, 3D printing, Fusion 360, SolidWorks, Onshape, and general fabrication
- Helped raise over $15,000 for Team 2059, contributing to two NC District Competition wins
- Team sites: Team 900, Team 2059
Thanks for looking at my profile! Connect with me or let me know if you'd like to work on a project!